This R package performs spectral analysis of multivariate spatial point patterns observed in a rectangular region (Ding et al., 2025). By specifying a parametric model for the intensity, it extracts the frequency‐domain characteristics of your data by smoothing the periodogram. Key features of this tool include spectral density estimation, cross-validated bandwidth selection, spectrum visualization, and coherence analysis.
Installation
You can install the development version of SpecPP from GitHub with:
# install.packages("pak")
pak::pak("qwding101/SpecPP")FAQ
Q: What is a (spatial) point pattern?
Spatial point pattern is a collection of locations of some event (observations) of interest on the space, such as earthquakes, crimes, disease outbreaks, traffic accidents, rental housing spots, and geotagged photos. A point pattern is multivariate if each event belongs to one of the finite set of categories. For an introduction to point pattern analysis, please refer to link 1 or link 2.
Q: What is the main advantage or unique contribution of SpecPP?
Traditional tools for point pattern analysis (in both spatial and frequency domains) often rely on the assumption of homogeneity, which is stringent for applications. Point pattern data in real life usually exhibits inhomogeneous behavior, that is, the locations of events are not scattered uniformly in space. In this package, we developed a frequency-domain method for analyzing inhomogeneous point patterns, which extends spectral techniques to multivariate and nonstationary settings. It also provides computational advantages, especially when there is a large number or type of events in the data.
Q: How to use SpecPP?
This tutorial demonstrates how to use SpecPP to analyze multivariate point pattern data.
References
- Ding, Q. W., Yang, J., & Shin, J. (2025). Pseudo-spectra of multivariate inhomogeneous spatial point processes. arXiv preprint arXiv:2502.09948.
- Yang, J., & Guan, Y. (2024). Fourier analysis of spatial point processes. arXiv preprint arXiv:2401.06403.